Abstract
Subsequent presentation of a previously viewed object results in improvement in behavioural response measures and also a decrease in brain activation as measured with fMRI. The suppressed BOLD response with primed stimuli has been related to suppressed firing of single neurons. Alternatively, suppressed BOLD response may be related to changes in the temporal pattern of firing of neurons. We suggest that primed objects produce activity that is shifted in time toward stimulus onset. Similar to an activity profile that is suppressed in magnitude, an activity profile that is shifted in time toward stimulus onset without changing peak activity produces less cumulative activity. A prediction based on the shifted profile is that primed objects should produce more activation during the pre-recognition interval. We measured BOLD responses during the pre-recognition interval by presenting masked stimuli in an event-related fMRI experiment. Stimuli were low-contrast images of familiar objects embedded in noise; they were flashed for 80 ms and then masked. Participants rated their level of recognition confidence using a 4-button response. Stimuli that were judged as completely unrecognizable were considered to produce activation that reflected only pre-recognition processes. For these unrecognized objects, primed stimuli produced more activation than non-primed stimuli. This finding is consistent with the hypothesized shift of activity in time, but not consistent with a general suppression mechanism. The findings are discussed in terms of an accumulation of neural activity during the pre-recognition interval that represents an accumulation of evidence for recognition.
Supported by NSF and JSMF (IG) and CIHR (TWJ)